Study On The Modeling And Optimization Of Stochastic Demand Vehicle Routing Problem With Simultaneous Delivery And Pick-Up

With the shortage of resources and environment pollution problem getting worse, ourgovernment pays more attention to the reuse of the used products, and many domestic enterprisesdevelop the reverse logistics system. In order to avoid the waste of vehicle transportation capacitycaused by the separate implementation logistics or reverse logistics, and to reduce the transportationcost, the enterprises often consider implementing the new products distribution and the used productsrecycling at the same time, thus formed vehicle routing problem with simultaneous delivery andpick-up(which is abbreviated VRPSDP). In existing research of VRPSDP, it is described that thecustomer’s demands for pick-up and delivery are known. However, the pickup demand of thecustomers is usually stochastic in real world, which formed the stochastic demand vehicle routingproblem with simultaneous delivery and pick-up(which is abbreviated SDVRPSDP). Due to randomfactors, how to solve the SDVRPSDP becomes very complex. Therefore, the study on the model andthe algorithm of SDVRPSDP becomes a problem which the reverse logistics enterprises have to face.The stochastic demand vehicle routing problem with simultaneous delivery and pickup is studiesin this paper. The main contents are listed below:Firstly, the research background and research signification of VRPSDP are introduced. Itscurrent researches are analyzed and some existing problems are pointed out. The main contributes andcreative points of this paper are elaborated and the research framework is presented.Secondly, the general definition and classification of VRPSDP are described in detail. Theclassical algorithms to solve the problem are analyzed. The Ant Colony System which is used in thispaper are introduced that lays a theoretical foundation for this research.Thirdly, the stochastic demand vehicle routing problem with simultaneous delivery and pick-upis discussed and a mixed integer programming model is constructed. In order to solve this problemeffectively, an Ant Colony System algorithm combining with Route Simulation Method and LocalSearch Method (which is abbreviated ACS-RSM) is proposed. The availability and effectiveness ofthis algorithm is proved by the case simulation in which the new algorithm and other algorithms areused.Fourthly, the stochastic demand vehicle routing problem with simultaneous delivery and pick-upand time window is analyzed and a mixed integer programming model is constructed. Considering thecharacteristics that the vehicles need to service the customers in the given time, an improvedACS-RSM algorithm is established, in which the route construction rules are redesign. The problem issolved successfully.